**3. Results**

#### *3.1. Analyses Applying the Flow Cytometry*

The flow cytometry analysis allowed evaluation of the total number of cells in each sample as well as live versus dead cells. This analysis was used to answer the question whether the number of bacterial cells is important in the evaluation of fecal microbiota and can be used as an estimator of a "good" or "bad" donor. The performed analysis showed that there were no statistically significant differences between cell numbers in the fecal microbiota suspensions prepared from the donors' stools (see Figure 2a). The summarized average cell count of all samples was equal to 1.664 × 10<sup>10</sup> cells/mL, (±0.913 × 10<sup>10</sup> cells/mL). Looking at numbers of live, dead and unknown cells each day for each donor, we noted discrete differences in the number of cells per mL of stool suspension; however, this was not statistically significant (donor A: mean 1.627 × 1010, ±1.015 × 1010; donor B: mean 1.632 × 1010, ±0.852 × 1010; donor C: mean 1.732 × 1010, ±0.96 ×<sup>10</sup>10). As shown in Figure 2b (right column), the percentages of all fractions of cells, i.e., alive, dead, unknown (not stained with one of the reagents) and SYTO9-PI- (subgroup of "unknown", not stained by both reagents considered as "double negative") showed relatively stable numbers in each stool sample per day and per donor. The noticeable domination of the alive cells was observed (Figure 2b).

**Figure 2.** Cytometry cell count charts. (**a**) Total cell count per donor summed over all samples. (**b**) Charts depicting variability in collected samples per day per donor. The first column shows absolute counts of cells classified as alive, dead or unknown. The second column shows cell counts as a percentage of the total number of cells counted in a given sample. (**c**) Two charts showing the variability of cells classified as a subgroup of unknown clusters: SYTO9-, PI-. The percentage was calculated versus the unknown group cell count.

In the next step, we specifically focused on alive cells, searching for if the number of this group of bacterial cells is important for the evaluation of fecal microbiota. As we observed, there were no significant differences in the viability of cells for each donor, and alive cells accounted for similar average percentages (Table S1).

#### *3.2. Cultivation of Stool Microbiota—Classical Microbiological Evaluation*

A complex cultivation experiment was performed to evaluate whether this technique can reveal culturable bacterial indicators for "good" versus "bad" stool donors. In total, 104 species representing 36 genera were found. The summarized bacterial species composition for each donor is indicated (Figure 3). Presentation of data in the form of a Venn diagram enabled us to indicate bacterial species that were characteristic of each donor, species that were shared by two donors and, finally, species that created a core microbiota and were found in all analyzed donors (Figure 3). Clearly, we can see that donor C, being a regular stool donor, is characterized by the largest number of cultivable species (64) obtained from his stool. Samples from other donors had lower numbers of cultivable species (48 and 56, respectively). Moreover, in the stool samples collected from donor C, the largest number of unique species (29) was found. Interestingly, the cultivable core microbiota was composed of only 16 species.

**Figure 3.** A three-set Venn diagram constructed based on the data from a classical microbiology approach. Identified genera are as follows: *Acidaminococcus*(*A. intestini*), *Arthrobacter*(*A. histidinolovorans*, *A. kerguelensis*), *Azoarcus* (*A. indigens*), *Bacillus* (*B. cereus*, *B. flexus*, *B. fordii*, *B. licheniformis*, *B. pumilus*, *B. safensis*), *Bacteroides*(*B. caccae*, *B. cellulosilyticus*, *B. coprocola*, *B. coprophilus*, *B. eggerthii*, *B. faecis*, *B. fragilis*, *B. massiliensis*, *B. nordi*, *B. ovatus*, *B. plebeius*, *B. uniformis*, *B. vulgatus*), *Bifidobacterium* (*B. adolescentis*, *B. bifidum*, *B. longum*, *B. pseudocatenulatum*, *B. ruminantium*), *Brevibacterium* (*B. casei*), *Brevundimonas* (*B. diminuta*, *B. vesicularis*), *Clostridium* (*C. beijerinckii*, *C. citroniae*, *C. innocuum*, *C. perfringens*, *C. symbiosum*, *C. tertium*, *C. thermopalmarium*), *Collinsella* (*C. aerofaciens*), *Coprobacillus* (*C. cateniformis*), *Corynebacterium* (*C. amycolatum*, *C. aurimucosum*, *C. minutissimum*), *Enterobacter* (*E. cloacae*, *E. kobei*), *Enterococcus* (*E. avium*, *E. casseliflavus*, *E. durans*, *E. faecalis*, *E. faecium*, *E. mundtii*, *E. thailandica*), *Escherichia* (*E. coli*), *Exiguobacterium* (*E. auranticum*), *Gordonia* (*G. rubripertincta*), *Klebsiella* (*K. oxytoca*, *K. pneumoniae*, *K. variicola*), *Kocuria* (*K. marina*, *K. varians*), *Lactobacillus* (*L. acidophilus*, *L. coleohominis*, *L. delbrueckii*, *L. fermentum*, *L. gasseri*, *L. jensenii*, *L. kefiri*, *L. paracasei*, *L. plantarum*, *L. salivarius*), *Lactococcus* (*L. garvieae*, *L. lactis*), *Lysinibacillus* (*L. boronitolerans*, *L. fusiformis*), *Microbacterium* (*M. aurum*, *M. lacticum*, *M. paraoxydans*), *Myroides* (*M. odoratimimus*), *Parabacteroides* (*P. distasonis*), *Penicillium* (*P. brevicompactum*), *Pseudomonas* (*P. alcaligenes*, *P. monteilii*, *P. putida*, *P. stutzerii*), *Rothia* (*R. dentocariosa*, *R. mucilaginosa*), *Slackia* (*S. heliotrinireducens*), *Sphingobacterium* (*S. mizutaii*), *Staphylococcus* (*S. epidermidis*, *S. hominis*), *Streptococcus* (*S. agalactiae*, *S. anginosus*, *S. constelatus*, *S. gallolyticus*, *S. mitis*, *S. oralis*, *S. parasanguinis*, *S. salivarius*, *S. vestibularis*), *Stretococcus* (*S. sanguinis*), *Sutterella* (*S. wadsworthensis*), *Veillonella* (*V. parvula*), *Weissella* (*W. confusa*, *W. viridescens*).

In the next step, we evaluated the presence of identified species over time, i.e., throughout 10 sampling days (Figure 4). It was shown that the most persistent species was *Escherichia coli*, being detected in all samples. Other species, such as *Enterococcus faecalis*, *Streprococcus parasanguinis*, *Bifidobacterium adolescentis*, *Enterococcus faecium* and *Streptococcus salivarius* were also detected in the majority of samples, ye<sup>t</sup> their persistence varied greatly between donors (Figure 4). It is noticeable that a plethora of bacterial species occurred only on individual days. This may be a consequence of the bias of this method or of a simple one-time variation depending on the food consumed. Therefore, when using conventional culturing as an evaluation strategy for the assessment of the quality of stool samples, it is important to repeat sampling from particular donors for several days. Single sampling can deliver non-representative and possibly false results.

### *3.3. Next-Generation Sequencing*

A total of 5,694,140 reads were obtained from Illumina MiSeq sequencing, with reads per sample ranging from 111,656 to 291,029. Quality control and merging of paired-end reads using the dada2 software package resulted in the retention of, on average, 47.74% (σ = 2.21) reads per sample (Table S2). Both the Nonpareil 3 and alpha rarefaction analysis (Qiime2 diversity plugin) showed sequencing depths close to 100%.

Merged reads subjected to further analyses were dereplicated into 9868 amplicon sequence variants (ASV), with the number of reads for ASV ranging from 1 to 10,098. Taxonomy assignment based on the Silva database (release 132), showed 97.75% of ASVs classified down to the genus level. Overall classification showed that 99.98% of all reads were bacterial, 0.01% archeal and less than 0.01% were unclassified. The bacterial ASVs represent 18 classes, with Bacteroidia and Clostridia in relative abundance, constituting averages of 49.9% and 40.0%, respectively. At a genus level, the most dominant taxa were *Bacteroides* and *Faecalibacterium*, with relative abundance in each sample no less than 35% and 11%, respectively. The data was not analyzed on species level because a single marker region does not allow this kind of resolution [34]. Figure 5 shows the abundances of each taxon identified in stool samples.

The Shannon diversity index along with Pielou's evenness index have been calculated for all of the samples, and the Kruskal–Wallis test was used to determine if there were any statistical di fferences between them. The Pielous evenness index for all of the samples was relatively high, ranging from 0.94 to 0.95, and no statistical di fferences between donors were detected. Interestingly, the Shannon index was similar for donors A and B, with its mean values equal to 10.11 and 10.02, while it was slightly, but significantly, higher for donor C—10.39 (*p* = 0.0191 for donor A versus C and *p* = 0.0005 for donor B versus C according to Kruskal–Wallis test, H value = 12.18; see Figure 6).

Given that the above statistical test for significant di fferences yielded two pairs, donor A versus C and donor B versus C, these pairs were subjected to ANCOM analysis. In the first case, ANCOM analysis highlighted two Gram-negative, obligatory anaerobe genera to be more abundant in samples from donor C; these were *Acidaminococcus* and *Paraprevotella*. The same analysis showed *Anaeroplasmatales* and *Gastranaerophilales* orders to be more abundant in samples from donor A.

An ANCOM analysis of the second pair (donor B versus C) highlighted *Anaeroplasma* as more abundant in donor B, along with *Holdermanella* genera and two, not well described, bacterial families—*Muribaculaceae* and *Puniceicoccaceae*, members of the *Bacteroidetes* and *Verrucomicrobia* phyla, respectively. As for taxa more abundant in the donor C microbiota, two members of the *Firmucutes* phylum were detected: *Lachnospiraceae* and *Dialister*.

**Figure 4.** Diagram showing the daily presence of the particular cultivable bacterial species in stool samples.

**Figure 5.** Heat map showing bacterial genera detected using amplicon sequencing (V3–V4 region of 16S rDNA). The summarized data for each donor are presented. The "others" group summarizes genera with individual abundances lower than 0.5% in any sample. Sequences unassigned at the genera taxonomy level were grouped and named "unassigned".

**Figure 6.** Boxplots showing selected biodiversity indices calculated for the data from the metabarcoding analysis. Kruskal-Wallis test was used to detect statistically significant differences. \*—*p*-value less than 0.01; \*\*\*—*p*-value less than 0.001.

As was done for classical culturing experiments, the stability of the intestinal microbiota over time was assessed. Although no statistically significant differences were detected, principal coordinates analysis (PCoA) on the Bray–Curtis dissimilarity index shows that the overall internal similarity of time-resolved samples from donor C was much higher than for other donors. Clustering the bacterial composition of feces in donor C indicates the most stable composition of intestinal microbiota over time (Figure 7).

Pearson correlation coefficients between the double negative group of cells (SYTO9-, PI-) and genera-level taxonomy data showed that relative abundance of *Anaeroplasma* is positively correlated with the double negative group per sample percentage (ρ = 0.6312), followed by *Sanguibacteroides* (ρ = 0.4592).

**Figure 7.** Principal coordinates analysis (PCoA) visualization. The PCoA was built using the Bray–Curtis dissimilarity index with day of sample collection as one axis.
